Source: redict
Version: 7.3.2+ds-1
Severity: wishlist

Hi! I'm opening this issue to track and discuss about dropping the jemalloc dependency from Debian's packaging of redict.

Required reading ("Proposed removal of jemalloc support"): <https://codeberg.org/redict/redict/issues/49> In short: the main reason why using jemalloc in redict makes sense is so that the "online memory defragmentation" feature can be used. Still, this only works when redict is compiled against its own fork of jemalloc. In Debian, we link redict to the system jemalloc library, which does not have the required redict-specific changes.

In order to simplify packaging and reduce redict's dependency tree, I propose using libc's standard malloc. Thoughts?

As a member of the team, I'll do this myself if others agree.

Bye!

Attachment: signature.asc
Description: PGP signature

Reply via email to